The 5v has low cross section areas. This is why I didn’t pull the sides in. The bottom should be fine though.

I know some of you want the 5v smaller I have cut it back but personally don’t want to take it back farther.

I’m okay pulling the sides in if they don’t have to be flat. (Gonna put bumps in edge in at least two spots.) sound okay?

Sphere478 wrote on 2022-12-07, 17:31:

The 5v has low cross section areas. This is why I didn’t pull the sides in. The bottom should be fine though.

you know what has low cross section area? 5V pin on the LDO 😀 you are over stressing

feipoa:

solder wick/desoldering braid + flux will fix that, but absolutely yes, moooaaar clearance!

Yeah, but a current path with a small choke point can still push more current than a wire as big as that choke point because of cooling and resistance.

I’ll think about it, but I personally don’t want the 5v any thinner.

If you look at latest draft I uploaded you can see I cut it back a lot.

I’ll try and pull the sides and bottom in tonight.

rasz_pl wrote on 2022-12-07, 17:12:

feipoa wrote on 2022-12-07, 03:34:

Removing those 4 tantalums from the centre region, which were 22 uF, 470 nF, 150 nF, 220 nF:
Noise on Vcc5 mostly disappears, or down to 48 mV and the noise on Vcc3 remains about the same.

fascinating, so they just acted like antenna?

Yeah, but there's a little bit of a gotcha. I'll report on it later. EDIT: It's an issue with tantalums for Cin on the LP regulator only, not MIC.

galanopu wrote on 2022-12-07, 13:38:

Coming soon... This was the request of someone who donated to me.
Results within the next few days. The rest in a video and then to the store.

Sounds a little like advertising. This forum isn't keen on that. I suggest you remove the part, "then to the store".

Lots of people around here with stores. 😀

Sphere478 wrote on 2022-12-07, 20:07:

Lots of people around here with stores. :-)

You need to take into account the whole context and probable purpose of the message, e.g. no comments in the thread previously, no critique on topic now, no mention of comparing values, etc. Considering these and the similarity to some of his other messages, I suspect the underlying purpose is to advertise, not community engagement.

Oh, Do you have junk PCB created by same company, can you dig away the via (the one with no hole) to see how big via is? The true via has a drilled hole bit larger and plated like old boards have?

I have a feelling these so called vias without the "hole" is tiny one?

Cheers,

Why all this toxicity? I think Sphere478 knows me and understands.
I am super busy, but I always replied when I was asked by him something technical.
This topic is just to hard to follow, let alone participate. It is more like a private chat, really.

[/quote]
no comments in the thread previously
[/quote]
Not true, I posted here before. I even offered my interest to help with PCB design but someone else was working on this, so I skipped.
I just do not like to work on other peoples work and maybe I am not that social to work with others.
As I said I only worked on this in the end, because of a donation request. I think this project has little financial value due to complexity.

[/quote]
no critique on topic now
[/quote]
Believe me, you do not want to hear my critique on the design choices here.
I can give you some tips as a professional PCB designer.
But based on your attitude maybe this is not a good idea.

[/quote]
no mention of comparing values, etc.
[/quote]
What does this even mean? Comparing what?
My design is completely different. You will have to wait for the results.

[/quote]
Considering these and the similarity to some of his other messages
[/quote]
For sure my posts are similar as I am the same person.
I make Youtube Videos and original retro hardware in a professional way.
And I am not the only one here, all these people are highly praised. I do not understand the attack.
I have never tried to sell here. I always just presented my results and maybe posted some of my videos.

[/quote]
I suspect the underlying purpose is to advertise, not community engagement.
[/quote]
Suspect? 🤣.... Advertise what? This is not even for sale now.
I am just here to give an early preview of my results.
As I said the final thing will be a video on my Youtube.
If you do not want to see my early results I will stop posting here.

@galanopu

I recognize the good intention in your message, just the wording feels a bit nuanced.
Better say it directly:
Hi there,
sharing an alternative approach ...
Linked CAD and BOM files.
I also offer pre-assembled items to save you the trouble. Here is the url.

There are other threads in the forums where people share links to purchasing options for assembled items or announce a price and ask who wants one.

I bought from you before and the experience was great (thank you for that), but others may not have the context.
It is ok really.

pshipkov wrote on 2022-12-08, 00:55:

@galanopu […]
Show full quote

@galanopu

I recognize the good intention in your message, just the wording feels a bit nuanced.
Better say it directly:
Hi there,
sharing an alternative approach ...
Linked CAD and BOM files.
I also offer pre-assembled items to save you the trouble. Here is the url.

There are other threads in the forums where people share links to purchasing options for assembled items or announce a price and ask who wants one.

I bought from you before and the experience was great (thank you for that), but others may not have the context.
It is ok really.

Hi pshipkov,
I surely remember you. Last time I could not service you in time due to how busy I am 😒

Thank you for the comment, however...
I do not do open hardware. So... "linked CAD and BOM files" just can not happen.
Some maybe get triggered by that, but I can not do much. I do this professionally.
As you said... "others may not have the context" you are right!
Still I can not phrase this much different when I do not do open stuff.

Yeah the only questionable part on my fist post was 3 words in the end of my comment.
"finally in the store" the meaning was just that these will be offered assembled (at some point).
I have never directly advertised here, or posted direct links to specific products.
Or mentioned anything about prices and all that. Or accepted orders from forums.
Usually I just post my videos here and most of these are not even about original hardware.

feipoa wrote on 2022-12-07, 23:42:

Sphere478 wrote on 2022-12-07, 20:07:

Lots of people around here with stores. 😀

You need to take into account the whole context and probable purpose of the message, e.g. no comments in the thread previously, no critique on topic now, no mention of comparing values, etc. Considering these and the similarity to some of his other messages, I suspect the underlying purpose is to advertise, not community engagement.

I suppose it kinda did lack a little tact 🤣 😀 but I have talked with him before nice guy. He made the little devices that allow you to overclock 486 boards If memory serves. Smart guy! 😀 I still need to get one of those! Feiopa, you might look into them.

Anyway, I’m working on it right now.

I once heard a story once about steve jobs and the first ipod. Who knows if it is true or not,

Engineers came to steve with a prototype for the brand new iPod and said here it is! The fruits of our labor behold it’s marvelous compact form! Steve said it needs to be smaller.😧 we can’t possibly make it smaller! They said. So he threw it in the fish tank and pointed at the bubbles.

I think you will like what I got coming 😀

Sphere478 wrote on 2022-12-08, 02:15:

Engineers came to steve with a prototype for the brand new iPod and said here it is! The fruits of our labor behold it’s marvelous compact form! Steve said it needs to be smaller.😧 we can’t possibly make it smaller! They said. So he threw it in the fish tank and pointed at the bubbles.
I think you will like what I got coming :-)

You are making us a boat? lol.
Edit: submarine?

galanopu wrote on 2022-12-07, 13:38:

Believe me, you do not want to hear my critique on the design choices here. I can give you some tips as a professional PCB designer

We welcome the critique, not the ego. If you have some relevant ideas for improvement, which will make a difference in its operational attributes, please enlighten. We are down to ~50 mV noise at 4 V (pre-clock doubling) and this matches the commercial unit. The device is fully functional at 80 MHz.

Here are the updates I've been wanting to present since yesterday. First, we are trying to mimic the noise and performance of the commercial Improve-It + Gainbery VRM interposer, which too shows a small amplitude, low frequency oscillation of 2.5 KHz. For these images, I will be using the 100 mV scale, rather than the 200 mV scale as previously shown.

I am leaving these 8x 100 nF 0805 caps on the PGA pins:

Now that we have cleared up the majority of the noise with those 8x 100 nF caps, I do see some additional improvement with the central caps. However, as noted on the previous page, the LP38503TSX-ADJ regulator does not like tantalums Cout, anywhere - opposite of the MIC29302WT. For the LP regulator, having 4x tantalums in the central region caused around 70 mV of additional noise to Vcc5 - not Vcc3. Also, having a tantalum at Cin1 caused 20 mV extra noise, which was periodic. The datasheet provides clarity on this with, If the source impedance is too high, the reactive component of the source may affect the control loop’s phase margin. To ensure proper loop operation, the ESR of the capacitor used for C IN must not exceed 0.5 Ω

With the MIC29302WT, having Cin1 as a ceramic maybe helped with Vcc3, but it was small and hard to quantify. Perhaps 4 mV. Cout1 or Cout centra being tantalum did not hurt the MIC's noise and is recommended based on information in the datasheet. In fact, it is stated, This capacitor need not be an expensive low ESR type: aluminum electrolytics are adequate. In fact, extremely low ESR capacitors may contribute to instability.

I asked pentiumspeed which ceramic cap values he recommended for the central region, but did not receive reply. So I decided to experiment with what ceramics I had in the 1210 and 1206 packages. I am using types X6S and X7R of sizes 10 uF, 100 nF, 220 nF, 10 uF. I tried these for the central region Cout separately from Cin, and then together on both the MIC and LP regulators. I will not present every graph because of the effort required, but I will present the before and after. I was concerned about the MIC using the ceramics, but perhaps they are far enough away from the regulator for the above warning not to come into play.

I also wanted to note that using 2x 10 uF ceramics at Cin1 made the noise a bit worse, not better.

Here is the MIC29302WT regulator with 8-100nf at the PGA pins, 1x Tantalum 10 uF Cout1, 1x MLCC 10 uF Cin1 and no central caps:

And here is the MIC29302WT regulator with 8-100nf at the PGA pins, 1x Tantalum 10 uF Cout1, 1x MLCC 10 uF Cin1 and central caps. Central caps are 10uF, 100nF, 220nF, 10uF:

Approximate Vcc3 noise before central caps: 88 mV
Approximate Vcc3 noise after central caps: 68 mV
Improvement in Vcc3 = 20 mV

Notice also that Vcc5 improved.
Approximate Vcc5 noise before central caps: 56 mV
Approximate Vcc5 noise after central caps: 40 mV
Improvement in Vcc5 = 16 mV

Following suite for the LP regulator, here is the LP38503TSX-ADJ regulator with 8-100nf at the PGA pins, 1x Tantalum 10 uF Cout1, 1x MLCC 10 uF Cin1 and no central caps:

And here is the LP38503TSX-ADJ regulator with 8-100nf at the PGA pins, 1x Tantalum 10 uF Cout1, 1x MLCC 10 uF Cin1 and central caps. Central caps are 10uF, 100nF, 220nF, 10uF:

Approximate Vcc3 noise before central caps: 68 mV
Approximate Vcc3 noise after central caps: 56 mV
Improvement in Vcc3 = 12 mV

Notice also that Vcc5 improved.
Approximate Vcc5 noise before central caps: 48 mV
Approximate Vcc5 noise after central caps: 44 mV
Improvement in Vcc5 = 4 mV

With the implemented caps, the IN and OUT noise in the feipoa-sphere replica unit using the LP38503TSX-ADJ now match that of the commercial product.

Note that I had tried ceramics on the central region before, but the reduction in noise was not apparent because the PGA 100 nF caps were not present. Also, when testing with the LP unit with central ceramics previously, there was a combination of tantalums that this reg doesn't like, only learned about later on. See notes in previous post.

While the LP regulator is OK with having cermaics on Cout, will it cause any issue with the MIC regulator? Obviously, the LP regulator is the preferred choice because of 20 mV less noise, but the MIC has the screw tab. Either way, both will work and it may come down to supplier availability.

EDIT: I had the MIC and LP images flipped. Been corrected.

Hey, any objection to reversing the wiring on the trimmer?

Was vrmsense vrmsense vcc

Wanna change it to vcc vcc vrmsense.

Should give same result if it is basically just a normal 3 lead potentiometer.

The modifications I am working on works a little better with shorter traces if I do this. Also less holes in vcc.

Do you want the central caps back in the design?

I had planned on leaving them out.

As requested in PM I made provisions to allow for 1210 capacitors.
There wasn’t a easy way to make a dual footprint because of the positioning and ground vias. So I just added silk so that they can be soldered to the same footprint as the tantalum without fear of shorting. Here are the two superimposed.

And here is how it will actually be. (Silk is the only thing I added)

Examining for errors, and stalling for more feedback.

I think I got all the bubbles out?

how does overclocking do with the new cap config?

I would like to see galanopu's interposer and details. can he post it again? can you guys let him? 😀 any chance you would show us schematics? or share the files? if not, that's fine 😀